训练单类别数据集时（使用rolabelImg制作300张图片）运行detect.py和val.py没有预测框。

[pangguanzhe344]

Search before asking

• I have searched the YOLOv5 issues and discussions and found no similar questions.

Question

Epoch gpu_mem box obj cls angle labels img_size
99/99 2.44G 0.1441 0.001176 0 0.002967 10 640
Class Images Labels P R mAP@.5 mAP@
all 30 0 0 0 0 0

100 epochs completed in 0.805 hours.
Optimizer stripped from runs/train/exp2/weights/last.pt, 15.4MB
Optimizer stripped from runs/train/exp2/weights/best.pt, 15.4MB

Validating runs/train/exp2/weights/best.pt...
Fusing layers...
Model Summary: 213 layers, 7498282 parameters, 0 gradients, 17.4 GFLOPs
Class Images Labels P R mAP@.5 mAP@
all 30 0 0 0 0 0
Results saved to runs/train/exp2

这是训练100轮的，每次都是Labels=0 P=0 R=0 ....是不是单类别的原因，还是其他问题？训练结果中的val_batch0_labels.jpg标出了label的框。
麻烦您解答。谢谢！

Additional

val: data=data/SKU-110K.yaml, weights=runs/train/exp2/weights/best.pt, batch_size=32, imgsz=640, conf_thres=0.001, iou_thres=0.6, task=val, device=, single_cls=False, augment=False, verbose=False, save_txt=False, save_hybrid=False, save_conf=False, save_json=False, project=runs/val, name=exp, exist_ok=False, half=False, dnn=False
YOLOv5 🚀 2021-11-18 torch 1.7.1+cu101 CUDA:0 (Tesla T4, 4308MiB)

Fusing layers...
Model Summary: 213 layers, 7498282 parameters, 0 gradients, 17.4 GFLOPs
val: Scanning 'data/labels/val.cache' images and labels... 30 found, 0 missing,
val: Scanning 'data/labels/val.cache' images and labels... 30 found, 0 missing,
val: Scanning 'data/labels/val.cache' images and labels... 30 found, 0 missing,
val: Scanning 'data/labels/val.cache' images and labels... 30 found, 0 missing,
val: Scanning 'data/labels/val.cache' images and labels... 30 found, 0 missing,
val: Scanning 'data/labels/val.cache' images and labels... 30 found, 0 missing,
val: Scanning 'data/labels/val.cache' images and labels... 30 found, 0 missing,
Class Images Labels P R mAP@.5 mAP@
all 30 0 0 0 0 0
Speed: 6.5ms pre-process, 16.8ms inference, 33.7ms NMS per image at shape (32, 3, 640, 640)
val: Scanning 'data/labels/val.cache' images and labels... 30 found, 0 missing,
Results saved to runs/val/exp

[acai66]

试试不要用单类别，自定义数据集.yaml 配置里nc改为2，name里加个None，像这样：

# Classes
nc: 2 
names: ['object', 'None']

[pangguanzhe344]

可以了！训练有结果了。在colab上训练的。感谢您的回答！跑通您的代码后，要努力看您的代码了！

[tkone2018]

@pangguanzhe344 你好，请问你训练的时候有预训练模型吗

[tkone2018]

@acai66 你好，可以分享一下demo模型吗，谢谢